<![CDATA[Advances in information technology have transformed the paradigm of accounting information systems and opened new areas for evolving fraud practices. This study investigates the relationship between fraud and accounting information systems through a systematic literature review of 10 journal articles published between 2023-2025. The study findings indicate that machine learning technologies, particularly ensemble learning and natural language processing, significantly contribute to detecting fraud in various accounting cycles. The revenue and receipts cycle and purchases and payables cycle are the most frequently exploited areas, while manipulation of journal entries through the general ledger is the most difficult form of fraud to detect. The digital competence and data science literacy of accounting personnel are proven to play an important role in detection effectiveness, with diagnostic skills serving as the main mediator, represents the most difficult form of fraud to detect. Digital competency and data science literacy of accounting personnel prove to be crucial factors in fraud detection effectiveness, with diagnostic skills serving as key mediators. The norms of segregation of duties and audit trails are also necessary in an increasingly digital environment, but they require modification to adapt to a computerized competitive environment that requires unique procedures compared to the manual era of the past. This study proposes five hypotheses, conducts further empirical research, and offers practical recommendations for decision-makers on enforcing preventive and detective strategies against fraudulent activities.]]>
Copyrights © 2026
